◆ Apache Airflow

Monitor and improve model performance

This is real work, not a feature someone invented — it comes from real job ads and real questions people asked. Below are four ready AI prompts: get it done, make it easy for the next person to say yes to, work out the right move when you are stuck, and stop it coming back.

4prompts
See the prompt↓All Apache Airflow tasks

The same task, four prompts

today's deadline · the next reviewer · the stuck moment · the pattern
AExecute — do the immediate taskSet up a daily job to re-train the customer churn prediction model using the latest three…+
Set up a daily job to re-train the customer churn prediction model using the latest three months of data, starting Monday, and alert the analytics team if the model's F1 score drops below 0.75.
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
BImprove — make it easier to acceptBefore deploying the re-training job for the churn model, ensure the new data is pre-processed…+
Before deploying the re-training job for the churn model, ensure the new data is pre-processed consistently with the original training data, and add a baseline comparison to the alert, so we only get notified if the performance degradation is significant and not just minor variance.
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
CDecide — diagnose the stuck momentThe customer churn model's F1 score has been fluctuating wildly for the past week, sometimes…+
The churn model's performance has been erratic, and I'm not sure if it's the model or the data.
The customer churn model's F1 score has been fluctuating wildly for the past week, sometimes dropping below the threshold, sometimes recovering. I'm not sure if it's a real model degradation, an issue with the incoming data, or a problem with the evaluation metric itself. The marketing team relies on this daily. I'm afraid of making a change that destabilizes it further. What's the most likely diagnosis, and what's the best next diagnostic step to pinpoint the root cause without causing more alarms?
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?
DBecome — change the patternI frequently struggle to distinguish between true model degradation and underlying data quality…+
I frequently struggle to distinguish between true model degradation and data quality issues.
I frequently struggle to distinguish between true model degradation and underlying data quality issues when performance alerts fire. This leads to wasted time chasing the wrong problem. What habit should I change to build more robust monitoring that helps me quickly differentiate between model and data problems, improving my diagnostic accuracy?
when the reply comes backPush once: ask it to sharpen the weakest part, and to say what it assumed. Helpful?

Questions people actually ask

honest answers, no sign-up

Every task here was seen in the real world. Someone doing the job named it, a real job ad asked for it, or a lot of people asked about it online.

If nothing real showed a task, it is not on the page. That is the whole rule.

They are the same job approached four ways, because what you need depends on where you are.

Get it done today. Make it easy for the next person to say yes to. Work out the right move when you are stuck. Learn the pattern so the job stops coming back.

For most of these jobs it can carry the heavy thinking - draft it, sort it, check it, rehearse it with you.

It cannot sit in your chair, take the blame when a number is wrong, or notice what nobody wrote down. Let it do the first 80%. Keep the last 20% that is truly yours.

No. Copy any prompt and paste it into the AI you already use. No account, no score, no wall in the way.

Any of them. The prompts describe the work rather than naming a product, so they are not tied to one assistant.

That is also why they keep working when you switch.

Change it freely. Every prompt is a starting line, not a rule.

Put in your real numbers, your real names and your real deadline. The more you make it yours, the better the answer comes back.

The tasks come from real job ads, published job data and the questions people ask in public forums.

The steps come from Apache Airflow's own documentation, with practitioner sources for the traps the manual does not mention.

Push once. Ask it to sharpen the weakest part and to say what it assumed.

Most wrong answers come from a missing detail rather than a bad prompt - tell it the thing it could not know.